Brendan Cully <brendan@kublai.com> [Fri, 28 Jul 2006 13:08:21 -0700] rev 2723
mq: do not hold a reference to repo in tags override
Instead, replace repo's class object with a subclass.
Christian Ebert <blacktrash@gmx.net> [Fri, 28 Jul 2006 22:17:32 +0200] rev 2722
patchbomb: fix generation of message-id when sending attachments
node was set to bin(node) before message-id was created
Vadim Gelfer <vadim.gelfer@gmail.com> [Fri, 28 Jul 2006 10:47:02 -0700] rev 2721
merge with crew.
Vadim Gelfer <vadim.gelfer@gmail.com> [Fri, 28 Jul 2006 10:46:41 -0700] rev 2720
mq: add qclone command
Vadim Gelfer <vadim.gelfer@gmail.com> [Fri, 28 Jul 2006 10:46:25 -0700] rev 2719
hg.py: add islocal() and defaultdest() functions, refactor
islocal tells if a repo or url is local.
defaultdest returns default path for clone if explicit path not given.
clone can now take repo or url as source
Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Fri, 28 Jul 2006 18:46:02 +0200] rev 2718
makes 'hg diff' diff the file in sorted order
Thomas Arendsen Hein <thomas@intevation.de> [Fri, 28 Jul 2006 09:01:13 +0200] rev 2717
merge with tonfa
Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Thu, 27 Jul 2006 19:26:01 +0200] rev 2716
reupdate the options after loading the repo
local .hgrc could possibly override them, command line take priority
Thomas Arendsen Hein <thomas@intevation.de> [Fri, 28 Jul 2006 09:00:59 +0200] rev 2715
merge with upstream
Matt Mackall <mpm@selenic.com> [Thu, 27 Jul 2006 15:20:35 -0500] rev 2714
Fix test permission
Thomas Arendsen Hein <thomas@intevation.de> [Fri, 28 Jul 2006 09:00:30 +0200] rev 2713
Fix test-mq-qnew-twice exit code and output.
Vadim Gelfer <vadim.gelfer@gmail.com> [Thu, 27 Jul 2006 18:24:59 -0700] rev 2712
mq: fix queue.apply to not call os.chdir()
call to os.chdir broke "hg -R dirname qpush".
instead pass -d (change directory) option to patch command.
Vadim Gelfer <vadim.gelfer@gmail.com> [Thu, 27 Jul 2006 16:41:59 -0700] rev 2711
mq: move many error messages to util.Abort
Vadim Gelfer <vadim.gelfer@gmail.com> [Thu, 27 Jul 2006 16:08:56 -0700] rev 2710
mq: do not allow to qnew a patch twice
Vadim Gelfer <vadim.gelfer@gmail.com> [Thu, 27 Jul 2006 15:53:08 -0700] rev 2709
run-tests.py: skip tests that should not run.
print message when any test is skipped.
count skipped tests.
Vadim Gelfer <vadim.gelfer@gmail.com> [Thu, 27 Jul 2006 15:31:04 -0700] rev 2708
make mq test more portable.
Vadim Gelfer <vadim.gelfer@gmail.com> [Thu, 27 Jul 2006 15:19:58 -0700] rev 2707
patchbomb: add content-disposition to make display inline and add filename
if rev being emailed has tag that ends in .patch or .diff then use that as
filename. else make up filename from name of repo.
Christian Ebert <blacktrash@gmx.net> [Thu, 27 Jul 2006 22:28:03 +0200] rev 2706
patchbomb: optionally send patches as inline attachments
Vadim Gelfer <vadim.gelfer@gmail.com> [Thu, 27 Jul 2006 12:36:17 -0700] rev 2705
merge with crew.
Christian Ebert <blacktrash@gmx.net> [Thu, 27 Jul 2006 21:25:39 +0200] rev 2704
make introductory message of patch series text/plain
Alexis S. L. Carvalho <alexis@cecm.usp.br> [Fri, 21 Jul 2006 01:51:17 -0300] rev 2703
run-tests.py: fix diff output when test-foo.out doesn't exist.
Previously, if test-foo.out doesn't exist and test-foo outputs
something, the first line of the diff will look like this:
-+changeset: 0:0acdaf898367
(note the leading "-")
Alexis S. L. Carvalho <alexis@cecm.usp.br> [Fri, 21 Jul 2006 01:51:17 -0300] rev 2702
Allow tests that end in .py and .bat
Revision 6ed46bad9530 disallowed tests that have a '.' in the name, but
that also disallows tests that end in .py and .bat, even though run-tests.py
has some code to special case them.
Thomas Arendsen Hein <thomas@intevation.de> [Thu, 27 Jul 2006 18:53:31 +0200] rev 2701
Merged backout head.
Thomas Arendsen Hein <thomas@intevation.de> [Thu, 27 Jul 2006 18:53:10 +0200] rev 2700
Backed out changeset 2ab464771b7d15c60014996d2853efc92cdb8465
- it always runs chmod
- uses chmod in an unsafe way
- if somebody does a clone of a repo, runs 'make tests', everything is fine,
so he can assume the original repo is fine, too, but it is not.